Neutralizing antibodies are considered to be essential for protection against many viral infections including those caused by retroviruses. Since the initial reports of neutralizing antibodies in HIV-infected individuals, it has become increasingly clear that high levels of these antibodies in serum correlate with better clinical outcome (3-5). These studies suggested that the identification of epitopes that elicit high titer neutralizing antibodies would be essential for vaccine development against HIV infection.
The primary receptor for the human immunodeficiency virus type 1 (HIV-1) is the CD4 molecule, found predominantly on the surface of T-lymphocytes. The binding of HIV-1 to CD4 occurs via the major viral envelope glycoprotein gp120 and initiates the viral infection process.
Current strategies for developing vaccines against infection by the human immunodeticiency virus have focused on eliciting antibodies against the viral envelope glycoprotein gp120 or its cell surface receptor CD4. Purified gp120 typically elicits type specific neutralizing antibodies that are reactive against epitopes that vary among virus isolates. This characteristic has hindered the use of gp120 as a vaccine.
CD4 has also been considered as a major candidate for development of a vaccine against HIV-1. Recent studies have demonstrated that sCD4 elicits HIV neutralizing antibodies in animals and prevents the spread of infection in SIV-infected rhesus monkeys (1). However, autoantibodies to CD4 may themselves create immune abnormalities in the immunized host if they interfere with normal T-cell functions. Neutralizing antibodies against gp120 are elicited in vivo in HIV-1-infected individuals and can be elicited in vitro using purified envelope glycoprotein. However, gp120 contains five hypervariable regions one of which, the V3 domain, is the principal neutralizing epitope. Hypervariability of this epitope among strains is a major obstacle for the generation of neutralizing antibodies effective against diverse strains of HIV-1. For these reasons it has been believed that vaccine strategies using either purified CD4 or gp120 present several disadvantages.
We have overcome the shortcomings of type specific anti gp120 antibodies and antibodies against CD4 by raising anti-HIV-1 neutralizing antibodies using as the immunogen a complex of gp120 chemically coupled to either soluble CD4 or to the mannose-specific lectin, succinyl concanavalin A (SC). We have found that these compounds induce similar conformational changes in gp120. The complexed gp120 appears to undergo a conformational change that presents an array of epitopes that were hidden on the uncomplexed glycoprotein (2). A portion of such epitopes elicit group-specific neutralizing antibodies, which are not strain limited like the type specific antibodies discussed above. We have discovered that the covalently bonded CD4-gp120 complexes are useful for raising neutralizing antibodies against various isolates of HIV-1 and against HIV-2.
The major research effort in the development of subunit vaccines against HIV has been directed toward the envelope glycoprotein of the virus. Inoculation of gp160 or gp120 into animals elicits neutralizing antibodies against HIV (3, 4). The principal neutralizing epitope on gp120 has been located between amino acids 306 and 326 in the third variable domain (V3 loop) of the protein (4). This epitope has been extensively studied by using both polyclonal and monoclonal antibodies (3, 4). In most cases antibodies directed to this region neutralize HIV-1 in an isolate specific manner although there is evidence that a weakly neutralizing species of anti-V3 loop antibodies can cross-react with diverse isolates (8). The reason for type specificity of anti-V3 loop antibodies is the extensive sequence variability among various isolates. Prolonged culturing of HIV-infected cells with type specific anti-V3 loop antibodies induces escape mutants resistant to neutralization (9).
In addition to the V3 loop, other neutralizing epicopes encompassing genetically conserved regions of the envelope have been identified (10, 11). However, immunization against these epitopes elicits polyclonal antisera with low neutralizing titers (12). For example, the CD4 binding region of gp120, encompassing a conserved region, elicits neutralizing antibodies against diverse isolates (13). However, this region is not normally an immunodominant epitope on the glycoprotein.
The interaction of gp120 with CD4 has been studied in considerable detail and regions of the molecules involved in complex formation have been determined (14-16). There are now several lines of evidence that interactions with CD4 induce conformational changes in gp120. First, recombinant soluble CD4 (sCD4) binding to gp120 increases the susceptibility of the V3 loop to monoclonal antibody binding and to digestion by exogenous proteinase (2). Second, sCD4 binding results in the dissociation of gp120 from the virus (17, 18). These conformational changes in gp120 are thought to facilitate the processes of virus attachment and fusion with the host cell membrane (2). Immunization with soluble CD4 and recombinant gp120, complexed by their natural affinity but not covalently bonded, resulted in the production of anti CD4 antibodies (31). Several murine monoclonal antibodies have been developed by immunization with mixtures of recombinant gp120 and sCD4 (31, 32). Antibodies raised in these studies were not strictly complex-specific and reacted with free gp120 or CD4; the neutralizing antibodies reacted with free sCD4, although they displayed various degrees of enhanced reactivity in the presence of gp120. The complexes used in these studies were unstable and comprised noncovalently bound gp120 and CD4.
A variety of N-linked carbohydrate structures of high mannose, complex and hybrid types present on the gp120 molecule may also play a role in the interaction of gp120 with host cell membranes (19-21). Indeed, a carbohydrate-mediated reactivity of gp120 has already been demonstrated with a serum lectin, known as mannose-binding protein, which has also been shown to inhibit HIV-1 infection of CD4+ cells (22). An additional carbohydrate-mediated interaction of gp120 has been shown with the endocytosis receptor of human macrophage membranes (21). It has been postulated that high affinity binding of accessible mannose residues on gp120 to the macrophage membrane may lead to virus uptake by the macrophage (21).
Recombinant soluble CD4 has been shown to inhibit HIV infection in vitro, mainly by competing with cell surface CD4. This observation has led to the possibility of using sCD4 for the therapy of HIV-infected individuals (23, 24). In addition, sCD4 has been used as an immunogen to block viral infection in animals. Treatment of SIV.sub.MAC -infected rhesus monkeys with sCD4 elicited not only an antibody response to human CD4 but also to monkey CD4. Coincident with the generation of such immunological responses, SIV could not be isolated from the PBL and bone marrow macrophages of these animals (1). A recent study with chimpanzees also demonstrated that human CD4 elicited anti-self CD4 antibody that inhibited HIV infection in vitro (25). Although immunization with sCD4 may be beneficial in blocking HIV infection, circulating antibody that recognizes self antigen may induce immune abnormality and dysfunction by binding to uninfected CD4+ cells. Nevertheless in theory anti-CD4 antibodies could be effective in blocking HIV infection provided they can disrupt virus attachment and entry without interfering with normal CD4 function. Ideally these antibodies should recognize CD4 epitopes that are present only after interaction with gp120.